endernon/wynnbuilder-forked-for-changes
1
0
Fork 0
forked from morphNon/wynnbuilder-idk
Code Pull requests Activity

Adv item search: rework entire parser to fix parsing issues

Browse source
This commit is contained in:
phantamanta44 2021-03-13 05:14:30 -06:00
parent 9c5877c061
commit 1693432462
5 changed files with 1389 additions and 459 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

818
expr_parser.js Normal file
View file

@ -0,0 +1,818 @@
/**
* See `expr_parser.md` for notes on the implementation of this parser!
*/
const ExprParser = (function() {
// buffer containing a list of tokens and a movable pointer into the list
class TokenBuf {
constructor(tokens) {
this.tokens = tokens;
this.index = 0;
}
get peek() {
return this.tokens[this.index];
}
consume(termName) {
if (this.peek.type !== termName) {
throw new Error(`Expected a ${termName}, but got a ${this.peek.type}!`)
}
const node = { type: 'term', token: this.peek };
++this.index;
return node;
}
}
// tokenize an expression string
function tokenize(exprStr) {
exprStr = exprStr.trim();
const tokens = [];
let col = 0;
function pushSymbol(sym) {
tokens.push({ type: sym });
col += sym.length;
}
while (col < exprStr.length) {
// parse fixed symbols, like operators and stuff
switch (exprStr[col]) {
case '(':
case ')':
case ',':
case '&':
case '|':
case '+':
case '-':
case '*':
case '/':
case '^':
case '=':
pushSymbol(exprStr[col]);
continue;
case '>':
pushSymbol(exprStr[col + 1] === '=' ? '>=' : '>');
continue;
case '<':
pushSymbol(exprStr[col + 1] === '=' ? '<=' : '<');
continue;
case '!':
pushSymbol(exprStr[col + 1] === '=' ? '!=' : '!');
continue;
case ' ': // ignore extra whitespace
++col;
continue;
}
if (exprStr.slice(col, col + 2) === "?=") {
pushSymbol("?=");
continue;
}
// parse a numeric literal
let m;
if ((m = /^\d+(?:\.\d*)?/.exec(exprStr.substring(col))) !== null) {
tokens.push({ type: 'nLit', value: parseFloat(m[0]) });
col += m[0].length;
continue;
}
// parse a string literal
if ((m = /^"([^"]+)"/.exec(exprStr.substring(col))) !== null) { // with double-quotes
tokens.push({ type: 'sLit', value: m[1] });
col += m[0].length;
continue;
}
if ((m = /^'([^']+)'/.exec(exprStr.substring(col))) !== null) { // with single-quotes
tokens.push({ type: 'sLit', value: m[1] });
col += m[0].length;
continue;
}
// parse an identifier or boolean literal
if ((m = /^\w[\w\d%]*/.exec(exprStr.substring(col))) !== null) {
switch (m[0]) {
case 'true':
tokens.push({ type: 'bLit', value: true });
col += 4;
continue;
case 'false':
tokens.push({ type: 'bLit', value: false });
col += 5;
continue;
}
tokens.push({ type: 'ident', id: m[0] });
col += m[0].length;
continue;
}
// if we reach here without successfully parsing a token, it's an error
throw new Error(`Could not parse character "${exprStr[col]}" at position ${col}`);
}
tokens.push({ type: 'eof' });
return new TokenBuf(tokens);
}
// parse an AST from a sequence of tokens
function takeExpr(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeConj(tokens));
return { type: 'nonterm', name: 'expr', prod: 0, children };
default:
throw new Error('Could not parse expr!');
}
}
function takeExprList(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeExpr(tokens));
children.push(takeExprList0(tokens));
return { type: 'nonterm', name: 'exprList', prod: 0, children };
default:
throw new Error('Could not parse exprList!');
}
}
function takeExprList0(tokens) {
const children = [];
switch (tokens.peek.type) {
case ')':
return { type: 'nonterm', name: 'exprList\'', prod: 1, children };
case ',':
children.push(tokens.consume(','));
children.push(takeExpr(tokens));
children.push(takeExprList0(tokens));
return { type: 'nonterm', name: 'exprList\'', prod: 0, children };
case 'eof':
return { type: 'nonterm', name: 'exprList\'', prod: 1, children };
default:
throw new Error('Could not parse exprList\'!');
}
}
function takeConj(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeDisj(tokens));
children.push(takeConj0(tokens));
return { type: 'nonterm', name: 'conj', prod: 0, children };
default:
throw new Error('Could not parse conj!');
}
}
function takeConj0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '&':
children.push(tokens.consume('&'));
children.push(takeDisj(tokens));
children.push(takeConj0(tokens));
return { type: 'nonterm', name: 'conj\'', prod: 0, children };
case ')':
case ',':
case 'eof':
return { type: 'nonterm', name: 'conj\'', prod: 1, children };
default:
throw new Error('Could not parse conj\'!');
}
}
function takeDisj(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeCmpEq(tokens));
children.push(takeDisj0(tokens));
return { type: 'nonterm', name: 'disj', prod: 0, children };
default:
throw new Error('Could not parse disj!');
}
}
function takeDisj0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '&':
case ')':
case ',':
return { type: 'nonterm', name: 'disj\'', prod: 1, children };
case '|':
children.push(tokens.consume('|'));
children.push(takeCmpEq(tokens));
children.push(takeDisj0(tokens));
return { type: 'nonterm', name: 'disj\'', prod: 0, children };
case 'eof':
return { type: 'nonterm', name: 'disj\'', prod: 1, children };
default:
throw new Error('Could not parse disj\'!');
}
}
function takeCmpEq(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeCmpRel(tokens));
children.push(takeCmpEq0(tokens));
return { type: 'nonterm', name: 'cmpEq', prod: 0, children };
default:
throw new Error('Could not parse cmpEq!');
}
}
function takeCmpEq0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!=':
children.push(tokens.consume('!='));
children.push(takeCmpRel(tokens));
children.push(takeCmpEq0(tokens));
return { type: 'nonterm', name: 'cmpEq\'', prod: 1, children };
case '&':
case ')':
case ',':
return { type: 'nonterm', name: 'cmpEq\'', prod: 3, children };
case '=':
children.push(tokens.consume('='));
children.push(takeCmpRel(tokens));
children.push(takeCmpEq0(tokens));
return { type: 'nonterm', name: 'cmpEq\'', prod: 0, children };
case '?=':
children.push(tokens.consume('?='));
children.push(takeCmpRel(tokens));
children.push(takeCmpEq0(tokens));
return { type: 'nonterm', name: 'cmpEq\'', prod: 2, children };
case '|':
case 'eof':
return { type: 'nonterm', name: 'cmpEq\'', prod: 3, children };
default:
throw new Error('Could not parse cmpEq\'!');
}
}
function takeCmpRel(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeSum(tokens));
children.push(takeCmpRel0(tokens));
return { type: 'nonterm', name: 'cmpRel', prod: 0, children };
default:
throw new Error('Could not parse cmpRel!');
}
}
function takeCmpRel0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!=':
case '&':
case ')':
case ',':
return { type: 'nonterm', name: 'cmpRel\'', prod: 4, children };
case '<':
children.push(tokens.consume('<'));
children.push(takeSum(tokens));
children.push(takeCmpRel0(tokens));
return { type: 'nonterm', name: 'cmpRel\'', prod: 1, children };
case '<=':
children.push(tokens.consume('<='));
children.push(takeSum(tokens));
children.push(takeCmpRel0(tokens));
return { type: 'nonterm', name: 'cmpRel\'', prod: 0, children };
case '=':
return { type: 'nonterm', name: 'cmpRel\'', prod: 4, children };
case '>':
children.push(tokens.consume('>'));
children.push(takeSum(tokens));
children.push(takeCmpRel0(tokens));
return { type: 'nonterm', name: 'cmpRel\'', prod: 2, children };
case '>=':
children.push(tokens.consume('>='));
children.push(takeSum(tokens));
children.push(takeCmpRel0(tokens));
return { type: 'nonterm', name: 'cmpRel\'', prod: 3, children };
case '?=':
case '|':
case 'eof':
return { type: 'nonterm', name: 'cmpRel\'', prod: 4, children };
default:
throw new Error('Could not parse cmpRel\'!');
}
}
function takeSum(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeProd(tokens));
children.push(takeSum0(tokens));
return { type: 'nonterm', name: 'sum', prod: 0, children };
default:
throw new Error('Could not parse sum!');
}
}
function takeSum0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!=':
case '&':
case ')':
return { type: 'nonterm', name: 'sum\'', prod: 2, children };
case '+':
children.push(tokens.consume('+'));
children.push(takeProd(tokens));
children.push(takeSum0(tokens));
return { type: 'nonterm', name: 'sum\'', prod: 0, children };
case ',':
return { type: 'nonterm', name: 'sum\'', prod: 2, children };
case '-':
children.push(tokens.consume('-'));
children.push(takeProd(tokens));
children.push(takeSum0(tokens));
return { type: 'nonterm', name: 'sum\'', prod: 1, children };
case '<':
case '<=':
case '=':
case '>':
case '>=':
case '?=':
case '|':
case 'eof':
return { type: 'nonterm', name: 'sum\'', prod: 2, children };
default:
throw new Error('Could not parse sum\'!');
}
}
function takeProd(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeExp(tokens));
children.push(takeProd0(tokens));
return { type: 'nonterm', name: 'prod', prod: 0, children };
default:
throw new Error('Could not parse prod!');
}
}
function takeProd0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!=':
case '&':
case ')':
return { type: 'nonterm', name: 'prod\'', prod: 2, children };
case '*':
children.push(tokens.consume('*'));
children.push(takeExp(tokens));
children.push(takeProd0(tokens));
return { type: 'nonterm', name: 'prod\'', prod: 0, children };
case '+':
case ',':
case '-':
return { type: 'nonterm', name: 'prod\'', prod: 2, children };
case '/':
children.push(tokens.consume('/'));
children.push(takeExp(tokens));
children.push(takeProd0(tokens));
return { type: 'nonterm', name: 'prod\'', prod: 1, children };
case '<':
case '<=':
case '=':
case '>':
case '>=':
case '?=':
case '|':
case 'eof':
return { type: 'nonterm', name: 'prod\'', prod: 2, children };
default:
throw new Error('Could not parse prod\'!');
}
}
function takeExp(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeUnary(tokens));
children.push(takeExp0(tokens));
return { type: 'nonterm', name: 'exp', prod: 0, children };
default:
throw new Error('Could not parse exp!');
}
}
function takeExp0(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!=':
case '&':
case ')':
case '*':
case '+':
case ',':
case '-':
case '/':
case '<':
case '<=':
case '=':
case '>':
case '>=':
case '?=':
return { type: 'nonterm', name: 'exp\'', prod: 1, children };
case '^':
children.push(tokens.consume('^'));
children.push(takeUnary(tokens));
children.push(takeExp0(tokens));
return { type: 'nonterm', name: 'exp\'', prod: 0, children };
case '|':
case 'eof':
return { type: 'nonterm', name: 'exp\'', prod: 1, children };
default:
throw new Error('Could not parse exp\'!');
}
}
function takeUnary(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
children.push(tokens.consume('!'));
children.push(takeUnary(tokens));
return { type: 'nonterm', name: 'unary', prod: 1, children };
case '(':
children.push(takePrim(tokens));
return { type: 'nonterm', name: 'unary', prod: 2, children };
case '-':
children.push(tokens.consume('-'));
children.push(takeUnary(tokens));
return { type: 'nonterm', name: 'unary', prod: 0, children };
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takePrim(tokens));
return { type: 'nonterm', name: 'unary', prod: 2, children };
default:
throw new Error('Could not parse unary!');
}
}
function takePrim(tokens) {
const children = [];
switch (tokens.peek.type) {
case '(':
children.push(tokens.consume('('));
children.push(takeExpr(tokens));
children.push(tokens.consume(')'));
return { type: 'nonterm', name: 'prim', prod: 4, children };
case 'bLit':
children.push(tokens.consume('bLit'));
return { type: 'nonterm', name: 'prim', prod: 1, children };
case 'ident':
children.push(tokens.consume('ident'));
children.push(takeIdentTail(tokens));
return { type: 'nonterm', name: 'prim', prod: 3, children };
case 'nLit':
children.push(tokens.consume('nLit'));
return { type: 'nonterm', name: 'prim', prod: 0, children };
case 'sLit':
children.push(tokens.consume('sLit'));
return { type: 'nonterm', name: 'prim', prod: 2, children };
default:
throw new Error('Could not parse prim!');
}
}
function takeIdentTail(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!=':
case '&':
return { type: 'nonterm', name: 'identTail', prod: 1, children };
case '(':
children.push(tokens.consume('('));
children.push(takeArgs(tokens));
children.push(tokens.consume(')'));
return { type: 'nonterm', name: 'identTail', prod: 0, children };
case ')':
case '*':
case '+':
case ',':
case '-':
case '/':
case '<':
case '<=':
case '=':
case '>':
case '>=':
case '?=':
case '^':
case '|':
case 'eof':
return { type: 'nonterm', name: 'identTail', prod: 1, children };
default:
throw new Error('Could not parse identTail!');
}
}
function takeArgs(tokens) {
const children = [];
switch (tokens.peek.type) {
case '!':
case '(':
children.push(takeExprList(tokens));
return { type: 'nonterm', name: 'args', prod: 0, children };
case ')':
return { type: 'nonterm', name: 'args', prod: 1, children };
case '-':
case 'bLit':
case 'ident':
case 'nLit':
case 'sLit':
children.push(takeExprList(tokens));
return { type: 'nonterm', name: 'args', prod: 0, children };
case 'eof':
return { type: 'nonterm', name: 'args', prod: 1, children };
default:
throw new Error('Could not parse args!');
}
}
// apply tree transformations to recover the nominal-grammar AST
function fixUp(ast) {
if (ast.type === 'term') {
return ast;
}
switch (ast.name) {
case 'exprList': // recover left-recursive structures
case 'conj':
case 'disj':
case 'exp':
return fixUpRightRecurse(ast, 1);
case 'sum':
case 'prod':
return fixUpRightRecurse(ast, 2);
case 'cmpEq':
return fixUpRightRecurse(ast, 3);
case 'cmpRel':
return fixUpRightRecurse(ast, 4);
case 'prim': // recover left-factored identifier things
switch (ast.prod) {
case 3:
switch (ast.children[1].prod) {
case 0: // function call
return {
type: 'nonterm', name: 'prim', prod: 3,
children: [fixUp(ast.children[0]), ...mapFixUp(ast.children[1].children)]
};
case 1: // just an identifier
return { type: 'nonterm', name: 'prim', prod: 4, children: [fixUp(ast.children[0])] };
}
break;
case 4:
return { ...ast, prod: 5, children: mapFixUp(ast.children) };
}
break;
}
return { ...ast, children: mapFixUp(ast.children) };
}
function mapFixUp(nodes) {
return nodes.map(chAst => fixUp(chAst));
}
function fixUpRightRecurse(ast, maxProd) {
let nomAst = { type: 'nonterm', name: ast.name, prod: maxProd, children: [fixUp(ast.children[0])] };
let tree = ast.children[1];
while (tree.prod < maxProd) {
nomAst = {
type: 'nonterm', name: ast.name, prod: tree.prod,
children: [nomAst, ...mapFixUp(tree.children.slice(0, tree.children.length - 1))]
};
tree = tree.children[tree.children.length - 1];
}
return nomAst;
}
// compile nominal AST into an item builder expression function
function translate(ast, builtInProps, builtInFuncs) {
function trans(ast) {
return translate(ast, builtInProps, builtInFuncs);
}
if (ast.type === 'term') {
switch (ast.token.type) {
case 'nLit':
return ast.token.value;
case 'bLit':
return ast.token.value;
case 'sLit':
return ast.token.value;
case 'ident':
return ast.token.value;
}
} else if (ast.type === 'nonterm') {
switch (ast.name) {
case 'expr':
return trans(ast.children[0]);
case 'exprList':
switch (ast.prod) {
case 0:
return [...ast.children[0], trans(ast.children[2])];
case 1:
return [trans(ast.children[0])];
}
break;
case 'conj':
switch (ast.prod) {
case 0:
return new ConjTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return trans(ast.children[0]);
}
break;
case 'disj':
switch (ast.prod) {
case 0:
return new DisjTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return trans(ast.children[0]);
}
break;
case 'cmpEq':
switch (ast.prod) {
case 0:
return new EqTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return new NeqTerm(trans(ast.children[0]), trans(ast.children[2]));
case 2:
return new ContainsTerm(trans(ast.children[0]), trans(ast.children[2]));
case 3:
return trans(ast.children[0]);
}
break;
case 'cmpRel':
switch (ast.prod) {
case 0:
return new LeqTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return new LtTerm(trans(ast.children[0]), trans(ast.children[2]));
case 2:
return new GtTerm(trans(ast.children[0]), trans(ast.children[2]));
case 3:
return new GeqTerm(trans(ast.children[0]), trans(ast.children[2]));
case 4:
return trans(ast.children[0]);
}
break;
case 'sum':
switch (ast.prod) {
case 0:
return new AddTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return new SubTerm(trans(ast.children[0]), trans(ast.children[2]));
case 2:
return trans(ast.children[0]);
}
break;
case 'prod':
switch (ast.prod) {
case 0:
return new MulTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return new DivTerm(trans(ast.children[0]), trans(ast.children[2]));
case 2:
return trans(ast.children[0]);
}
break;
case 'exp':
switch (ast.prod) {
case 0:
return new ExpTerm(trans(ast.children[0]), trans(ast.children[2]));
case 1:
return trans(ast.children[0]);
}
break;
case 'unary':
switch (ast.prod) {
case 0:
return new NegTerm(trans(ast.children[1]));
case 1:
return new InvTerm(trans(ast.children[1]));
case 2:
return trans(ast.children[0]);
}
break;
case 'prim':
switch (ast.prod) {
case 0:
return new NumLitTerm(ast.children[0].token.value);
case 1:
return new BoolLitTerm(ast.children[0].token.value);
case 2:
return new StrLitTerm(ast.children[0].token.value);
case 3:
const fn = builtInFuncs[ast.children[0].token.id];
if (!fn) {
throw new Error(`Unknown function: ${ast.children[0].token.id}`);
}
return new FnCallTerm(fn, trans(ast.children[2]));
case 4: {
const prop = builtInProps[ast.children[0].token.id];
if (!prop) {
throw new Error(`Unknown property: ${ast.children[0].token.id}`);
}
return new PropTerm(prop);
}
case 5:
return trans(ast.children[1]);
}
break;
case 'args':
switch (ast.prod) {
case 0:
return trans(ast.children[0]);
case 1:
return [];
}
break;
}
}
}
return class ExprParser {
constructor(builtInProps, builtInFuncs) {
this.builtInProps = builtInProps;
this.builtInFuncs = builtInFuncs;
}
parse(exprStr) {
const tokens = tokenize(exprStr);
if (tokens.tokens.length <= 1) {
return null;
}
const transAst = takeExpr(tokens, 0);
if (tokens.peek.type !== 'eof') {
throw new Error('Could not parse entire expression!');
}
const nomAst = fixUp(transAst);
return translate(nomAst, this.builtInProps, this.builtInFuncs);
}
};
})();

197
expr_parser.md Normal file
View file

@ -0,0 +1,197 @@
This file contains notes on how the parser implemented in `expr_parser.js` is designed.
The nominal grammar is as follows:
```
expr := conj
exprList := exprList "," expr
| expr
conj := conj "&" disj
| disj
disj := disj "|" cmpEq
| cmpEq
cmpEq := cmpEq "=" cmpRel
| cmpEq "!=" cmpRel
| cmpEq "?=" cmpRel
| cmpRel
cmpRel := cmpRel "<=" sum
| cmpRel "<" sum
| cmpRel ">" sum
| cmpRel ">=" sum
| sum
sum := sum "+" prod
| sum "-" prod
| prod
prod := prod "*" exp
| prod "/" exp
| exp
exp := exp "^" unary
| unary
unary := "-" unary
| "!" unary
| prim
prim := nLit
| bLit
| sLit
| ident "(" args ")"
| ident
| "(" expr ")"
args := exprList
| ε
```
In order to eliminate left-recursion, the above grammar is transformed into the following:
```
expr := conj
exprList := expr exprListTail
exprList' := "," expr exprList'
| ε
conj := disj conj'
conj' := "&" disj conj'
| ε
disj := cmpEq disj'
disj' := "|" cmpEq disj'
| ε
cmpEq := cmpRel cmpEq'
cmpEq' := "=" cmpRel cmpEq'
| "!=" cmpRel cmpEq'
| "?=" cmpRel cmpEq'
| ε
cmpRel := sum cmpRel'
cmpRel' := "<=" sum cmpRel'
| "<" sum cmpRel'
| ">" sum cmpRel'
| ">=" sum cmpRel'
| ε
sum := prod sum'
sum' := "+" prod sum'
| "-" prod sum'
| ε
prod := exp prod'
prod' := "*" exp prod'
| "/" exp prod'
| ε
exp := unary exp'
exp' := "^" unary exp'
| ε
unary := "-" unary
| "!" unary
| prim
prim := nLit
| bLit
| sLit
| ident identTail
| "(" expr ")"
identTail := "(" args ")"
| ε
args := exprList
| ε
```
The parser itself is a recursive-descent LL(1) parser.
To build the parsing rules, the following `FIRST` and `FOLLOW` sets are computed for the above grammar:
Nonterminal | FIRST
----------- | -----
`expr` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`exprList` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`exprList'` | `","`, `ε`
`conj` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`conj'` | `"&"`, `ε`
`disj` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`disj'` | `"\|"`, `ε`
`cmpEq` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`cmpEq'` | `"!="`, `"="`, `"?="`, `ε`
`cmpRel` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`cmpRel'` | `"<"`, `"<="`, `">"`, `">="`, `ε`
`sum` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`sum'` | `"+"`, `"-"`, `ε`
`prod` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`prod'` | `"*"`, `"/"`, `ε`
`exp` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`exp'` | `"^"`, `ε`
`unary` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`
`prim` | `"("`, `bLit`, `ident`, `nLit`, `sLit`
`identTail` | `"("`, `ε`
`args` | `"!"`, `"("`, `"-"`, `bLit`, `ident`, `nLit`, `sLit`, `ε`
Nonterminal | FOLLOW
----------- | ------
`expr` | `")"`, `","`, `$`
`exprList` | `")"`
`exprList'` | `")"`
`conj` | `")"`, `","`
`conj'` | `")"`, `","`
`disj` | `"&"`, `")"`, `","`
`disj'` | `"&"`, `")"`, `","`
`cmpEq` | `"&"`, `")"`, `","`, `"\|"`
`cmpEq'` | `"&"`, `")"`, `","`, `"\|"`
`cmpRel` | `"!="`, `"&"`, `")"`, `","`, `"="`, `"?="`, `"\|"`
`cmpRel'` | `"!="`, `"&"`, `")"`, `","`, `"="`, `"?="`, `"\|"`
`sum` | `"!="`, `"&"`, `")"`, `","`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"\|"`
`sum'` | `"!="`, `"&"`, `")"`, `","`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"\|"`
`prod` | `"!="`, `"&"`, `")"`, `"+"`, `","`, `"-"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"\|"`
`prod'` | `"!="`, `"&"`, `")"`, `"+"`, `","`, `"-"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"\|"`
`exp` | `"!="`, `"&"`, `")"`, `"*"`, `"+"`, `","`, `"-"`, `"/"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"\|"`
`exp'` | `"!="`, `"&"`, `")"`, `"*"`, `"+"`, `","`, `"-"`, `"/"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"\|"`
`unary` | `"!="`, `"&"`, `")"`, `"*"`, `"+"`, `","`, `"-"`, `"/"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"^"`, `"\|"`
`prim` | `"!="`, `"&"`, `")"`, `"*"`, `"+"`, `","`, `"-"`, `"/"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"^"`, `"\|"`
`identTail` | `"!="`, `"&"`, `")"`, `"*"`, `"+"`, `","`, `"-"`, `"/"`, `"<"`, `"<="`, `"="`, `">"`, `">="`, `"?="`, `"^"`, `"\|"`
`args` | `")"`
Then, the parsing rule table is as follows, where each table entry refers to the production number for the given nonterminal:
Nonterminal | "!" | "!=" | "&" | "(" | ")" | "*" | "+" | "," | "-" | "/" | "<" | "<=" | "=" | ">" | ">=" | "?=" | "^" | "&#124;" | bLit | ident | nLit | sLit | $
----------- | --- | ---- | --- | --- | --- | --- | --- | --- | --- | --- | --- | ---- | --- | --- | ---- | ---- | --- | --- | ---- | ----- | ---- | ---- | ---
expr | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
exprList | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
exprList' | - | - | - | - | 1 | - | - | 0 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 1
conj | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
conj' | - | - | 0 | - | 1 | - | - | 1 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 1
disj | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
disj' | - | - | 1 | - | 1 | - | - | 1 | - | - | - | - | - | - | - | - | - | 0 | - | - | - | - | 1
cmpEq | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
cmpEq' | - | 1 | 3 | - | 3 | - | - | 3 | - | - | - | - | 0 | - | - | 2 | - | 3 | - | - | - | - | 3
cmpRel | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
cmpRel' | - | 4 | 4 | - | 4 | - | - | 4 | - | - | 1 | 0 | 4 | 2 | 3 | 4 | - | 4 | - | - | - | - | 4
sum | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
sum' | - | 2 | 2 | - | 2 | - | 0 | 2 | 1 | - | 2 | 2 | 2 | 2 | 2 | 2 | - | 2 | - | - | - | - | 2
prod | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
prod' | - | 2 | 2 | - | 2 | 0 | 2 | 2 | 2 | 1 | 2 | 2 | 2 | 2 | 2 | 2 | - | 2 | - | - | - | - | 2
exp | 0 | - | - | 0 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | -
exp' | - | 1 | 1 | - | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 0 | 1 | - | - | - | - | 1
unary | 1 | - | - | 2 | - | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 2 | 2 | 2 | 2 | -
prim | - | - | - | 4 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | 1 | 3 | 0 | 2 | -
identTail | - | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | - | - | - | - | 1
args | 0 | - | - | 0 | 1 | - | - | - | 0 | - | - | - | - | - | - | - | - | - | 0 | 0 | 0 | 0 | 1

1
items_2.html
View file

@ -78,6 +78,7 @@
<script type="text/javascript" src="damage_calc.js"></script>
<script type="text/javascript" src="display.js"></script>
<script type="text/javascript" src="query_2.js"></script>
<script type="text/javascript" src="expr_parser.js"></script>
<script type="text/javascript" src="load.js"></script>
<script type="text/javascript" src="items_2.js"></script>
</body>

27
items_2.js
View file

@ -21,6 +21,7 @@ class ExprField {
} catch (e) {
this.errorText.innerText = e.message;
this.output = null;
console.error(e);
}
return true;
}
@ -38,6 +39,8 @@ function compareLexico(ia, keysA, ib, keysB) {
if (aKey > bKey) return 1;
break;
case 'number': // sort numeric stuff in reverse order
aKey = isNaN(aKey) ? 0 : aKey;
bKey = isNaN(bKey) ? 0 : bKey;
if (aKey < bKey) return 1;
if (aKey > bKey) return -1;
break;
@ -74,17 +77,23 @@ function init() {
itemEntries.push(itemElem);
}
// create the expression parser
const exprParser = new ExprParser(itemQueryProps, itemQueryFuncs);
// the two search query input boxes
const searchFilterField = new ExprField('search-filter-field', 'search-filter-error', function(exprStr) {
const expr = compileQueryExpr(exprStr);
return expr !== null ? expr : (i, ie) => true;
const expr = exprParser.parse(exprStr);
return expr !== null ? expr : new BoolLitTerm(true);
});
const searchSortField = new ExprField('search-sort-field', 'search-sort-error', function(exprStr) {
const subExprs = exprStr.split(';').map(compileQueryExpr).filter(f => f != null);
return function(i, ie) {
const sortKeys = [];
for (let k = 0; k < subExprs.length; k++) sortKeys.push(subExprs[k](i, ie));
return sortKeys;
const subExprs = exprStr.split(';').map(e => exprParser.parse(e)).filter(f => f != null);
return {
type: 'array',
resolve(i, ie) {
const sortKeys = [];
for (let k = 0; k < subExprs.length; k++) sortKeys.push(subExprs[k].resolve(i, ie));
return sortKeys;
}
};
});
@ -108,8 +117,8 @@ function init() {
try {
for (let i = 0; i < searchDb.length; i++) {
const item = searchDb[i][0], itemExp = searchDb[i][1];
if (checkBool(searchFilterField.output(item, itemExp))) {
searchResults.push({ item, itemExp, sortKeys: searchSortField.output(item, itemExp) });
if (checkBool(searchFilterField.output.resolve(item, itemExp))) {
searchResults.push({ item, itemExp, sortKeys: searchSortField.output.resolve(item, itemExp) });
}
}
} catch (e) {

805
query_2.js
View file

@ -1,61 +1,4 @@
/*
* disj := conj "|" disj
* | conj
*
* conj := cmp "&" conj
* | cmpEq
*
* cmpEq := cmpRel "=" cmpEq
* | cmpRel "?=" prim
* | cmpRel "!=" cmpEq
*
* cmpRel := sum "<=" cmpRel
* | sum "<" cmpRel
* | sum ">" cmpRel
* | sum ">=" cmpRel
* | sum
*
* sum := prod "+" sum
* | prod "-" sum
* | prod
*
* prod := exp "*" prod
* | exp "/" prod
* | exp
*
* exp := unary "^" exp
* | unary
*
* unary := "-" unary
* | "!" unary
* | prim
*
* prim := nLit
* | bLit
* | sLit
* | ident "(" [disj ["," disj...]] ")"
* | ident
* | "(" disj ")"
*/
// a list of tokens indexed by a single pointer
class TokenList {
constructor(tokens) {
this.tokens = tokens;
this.ptr = 0;
}
get here() {
if (this.ptr >= this.tokens.length) throw new Error('Reached end of expression');
return this.tokens[this.ptr];
}
advance(steps = 1) {
this.ptr = Math.min(this.ptr + steps, this.tokens.length);
}
}
// type casts
// dynamic type casts
function checkBool(v) {
if (typeof v !== 'boolean') throw new Error(`Expected boolean, but got ${typeof v}`);
return v;
@ -71,67 +14,79 @@ function checkStr(v) {
return v;
}
function checkComparable(v) {
if (typeof v === 'boolean') throw new Error('Boolean is not comparable');
return v;
}
// properties of items that can be looked up
// each entry is a function `(item, extended item) -> value`
const itemQueryProps = (function() {
const props = {};
function prop(names, getProp) {
function prop(names, type, resolve) {
if (Array.isArray(names)) {
for (name of names) {
props[name] = getProp;
props[name] = { type, resolve };
}
} else {
props[names] = getProp;
props[names] = { type, resolve };
}
}
function maxId(names, idKey) {
prop(names, (i, ie) => ie.get('maxRolls').get(idKey) || 0);
prop(names, 'number', (i, ie) => ie.get('maxRolls').get(idKey) || 0);
}
function minId(names, idKey) {
prop(names, (i, ie) => ie.get('minRolls').get(idKey) || 0);
prop(names, 'number', (i, ie) => ie.get('minRolls').get(idKey) || 0);
}
function rangeAvg(names, getProp) {
prop(names, (i, ie) => {
prop(names, 'number', (i, ie) => {
const range = getProp(i, ie);
if (!range) return 0;
const ndx = range.indexOf('-');
return (parseInt(range.substring(0, ndx), 10) + parseInt(range.substring(ndx + 1), 10)) / 2;
});
}
function map(names, comps, f) {
return prop(names, (i, ie) => {
function map(names, comps, outType, f) {
return prop(names, outType, (i, ie) => {
const args = [];
for (let k = 0; k < comps.length; k++) args.push(comps[k](i, ie));
return f.apply(null, args);
});
}
function sum(names, ...comps) {
return map(names, comps, (...summands) => {
return map(names, comps, 'number', (...summands) => {
let total = 0;
for (let i = 0; i < summands.length; i++) total += summands[i];
return total;
});
}
prop('name', (i, ie) => i.displayName || i.name);
prop('type', (i, ie) => i.type);
prop(['cat', 'category'], (i, ie) => i.category);
prop('name', 'string', (i, ie) => i.displayName || i.name);
prop('type', 'string', (i, ie) => i.type);
prop(['cat', 'category'], 'string', (i, ie) => i.category);
const tierIndices = { Normal: 0, Unique: 1, Set: 2, Rare: 3, Legendary: 4, Fabled: 5, Mythic: 6 };
prop(['rarityname', 'raritystr', 'tiername', 'tierstr'], (i, ie) => i.tier);
prop(['rarity', 'tier'], (i, ie) => tierIndices[i.tier]);
prop(['rarityname', 'raritystr', 'tiername', 'tierstr'], 'string', (i, ie) => i.tier);
prop(['rarity', 'tier'], 'string', (i, ie) => tierIndices[i.tier]);
prop(['level', 'lvl', 'combatlevel', 'combatlvl'], (i, ie) => i.lvl);
prop(['strmin', 'strreq'], (i, ie) => i.strReq);
prop(['dexmin', 'dexreq'], (i, ie) => i.dexReq);
prop(['intmin', 'intreq'], (i, ie) => i.intReq);
prop(['defmin', 'defreq'], (i, ie) => i.defReq);
prop(['agimin', 'agireq'], (i, ie) => i.agiReq);
prop(['level', 'lvl', 'combatlevel', 'combatlvl'], 'number', (i, ie) => i.lvl);
prop(['strmin', 'strreq'], 'number', (i, ie) => i.strReq);
prop(['dexmin', 'dexreq'], 'number', (i, ie) => i.dexReq);
prop(['intmin', 'intreq'], 'number', (i, ie) => i.intReq);
prop(['defmin', 'defreq'], 'number', (i, ie) => i.defReq);
prop(['agimin', 'agireq'], 'number', (i, ie) => i.agiReq);
sum(['summin', 'sumreq', 'totalmin', 'totalreq'], props.strmin, props.dexmin, props.intmin, props.defmin, props.agimin);
prop('str', (i, ie) => i.str);
prop('dex', (i, ie) => i.dex);
prop('int', (i, ie) => i.int);
prop('def', (i, ie) => i.def);
prop('agi', (i, ie) => i.agi);
prop('str', 'number', (i, ie) => i.str);
prop('dex', 'number', (i, ie) => i.dex);
prop('int', 'number', (i, ie) => i.int);
prop('def', 'number', (i, ie) => i.def);
prop('agi', 'number', (i, ie) => i.agi);
sum(['skillpoints', 'skillpts', 'attributes', 'attrs'], props.str, props.dex, props.int, props.def, props.agi);
rangeAvg(['neutraldmg', 'neutraldam', 'ndmg', 'ndam'], (i, ie) => i.nDam);
@ -155,15 +110,15 @@ const itemQueryProps = (function() {
maxId(['spellrawdmg', 'spellrawdam', 'spellneutraldmg', 'spellneutraldam', 'sdraw'], 'sdRaw');
const atkSpdIndices = { SUPER_SLOW: -3, VERY_SLOW: -2, SLOW: -1, NORMAL: 0, FAST: 1, VERY_FAST: 2, SUPER_FAST: 3 };
prop(['attackspeed', 'atkspd'], (i, ie) => i.atkSpd ? atkSpdIndices[i.atkSpd] : 0);
prop(['attackspeed', 'atkspd'], 'string', (i, ie) => i.atkSpd ? atkSpdIndices[i.atkSpd] : 0);
maxId(['bonusattackspeed', 'bonusatkspd', 'attackspeedid', 'atkspdid', 'atktier'], 'atkTier');
sum(['sumattackspeed', 'totalattackspeed', 'sumatkspd', 'totalatkspd', 'sumatktier', 'totalatktier'], props.atkspd, props.atktier);
prop(['earthdef', 'edef'], (i, ie) => i.eDef || 0);
prop(['thunderdef', 'tdef'], (i, ie) => i.tDef || 0);
prop(['waterdef', 'wdef'], (i, ie) => i.wDef || 0);
prop(['firedef', 'fdef'], (i, ie) => i.fDef || 0);
prop(['airdef', 'adef'], (i, ie) => i.aDef || 0);
prop(['earthdef', 'edef'], 'number', (i, ie) => i.eDef || 0);
prop(['thunderdef', 'tdef'], 'number', (i, ie) => i.tDef || 0);
prop(['waterdef', 'wdef'], 'number', (i, ie) => i.wDef || 0);
prop(['firedef', 'fdef'], 'number', (i, ie) => i.fDef || 0);
prop(['airdef', 'adef'], 'number', (i, ie) => i.aDef || 0);
sum(['sumdef', 'totaldef'], props.edef, props.tdef, props.wdef, props.fdef, props.adef);
maxId(['earthdef%', 'edef%', 'edefpct'], 'eDefPct');
@ -173,7 +128,7 @@ const itemQueryProps = (function() {
maxId(['airdef%', 'adef%', 'adefpct'], 'aDefPct');
sum(['sumdef%', 'totaldef%', 'sumdefpct', 'totaldefpct'], props.edefpct, props.tdefpct, props.wdefpct, props.fdefpct, props.adefpct);
prop(['health', 'hp'], (i, ie) => i.hp || 0);
prop(['health', 'hp'], 'number', (i, ie) => i.hp || 0);
maxId(['bonushealth', 'healthid', 'bonushp', 'hpid', 'hpbonus'], 'hpBonus');
sum(['sumhealth', 'sumhp', 'totalhealth', 'totalhp'], props.hp, props.hpid);
@ -207,397 +162,347 @@ const itemQueryProps = (function() {
maxId(['lootbonus', 'lb'], 'lb');
maxId(['xpbonus', 'xpb', 'xb'], 'xpb');
maxId(['stealing', 'esteal'], 'eSteal');
prop(['powderslots', 'powders', 'slots', 'sockets'], (i, ie) => i.slots || 0);
prop(['powderslots', 'powders', 'slots', 'sockets'], 'number', (i, ie) => i.slots || 0);
return props;
})();
// functions that can be called in query expressions
const itemQueryFuncs = {
max(args) {
if (args.length < 1) throw new Error('Not enough args to max()');
let runningMax = -Infinity;
for (let i = 0; i < args.length; i++) {
if (checkNum(args[i]) > runningMax) runningMax = args[i];
max: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to max()');
let runningMax = -Infinity;
for (let i = 0; i < args.length; i++) {
if (checkNum(args[i]) > runningMax) runningMax = args[i];
}
return runningMax;
}
return runningMax;
},
min(args) {
if (args.length < 1) throw new Error('Not enough args to min()');
let runningMin = Infinity;
for (let i = 0; i < args.length; i++) {
if (checkNum(args[i]) < runningMin) runningMin = args[i];
min: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to min()');
let runningMin = Infinity;
for (let i = 0; i < args.length; i++) {
if (checkNum(args[i]) < runningMin) runningMin = args[i];
}
return runningMin;
}
return runningMin;
},
floor(args) {
if (args.length < 1) throw new Error('Not enough args to floor()');
return Math.floor(checkNum(args[0]));
},
ceil(args) {
if (args.length < 1) throw new Error('Not enough args to ceil()');
return Math.ceil(checkNum(args[0]));
},
round(args) {
if (args.length < 1) throw new Error('Not enough args to ceil()');
return Math.round(checkNum(args[0]));
},
sqrt(args) {
if (args.length < 1) throw new Error('Not enough args to ceil()');
return Math.sqrt(checkNum(args[0]));
},
abs(args) {
if (args.length < 1) throw new Error('Not enough args to ceil()');
return Math.abs(checkNum(args[0]));
},
contains(args) {
if (args.length < 2) throw new Error('Not enough args to contains()');
return checkStr(args[0]).toLowerCase().includes(checkStr(args[1]).toLowerCase());
},
atkspdmod(args) {
if (args.length < 1) throw new Error('Not enough args to atkSpdMod()');
switch (checkNum(args[0])) {
case 2: return 3.1;
case 1: return 2.5;
case 0: return 2.05;
case -1: return 1.5;
case -2: return 0.83;
floor: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to floor()');
return Math.floor(checkNum(args[0]));
}
},
ceil: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to ceil()');
return Math.ceil(checkNum(args[0]));
}
},
round: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to round()');
return Math.round(checkNum(args[0]));
}
},
sqrt: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to sqrt()');
return Math.sqrt(checkNum(args[0]));
}
},
abs: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to abs()');
return Math.abs(checkNum(args[0]));
}
},
contains: {
type: 'boolean',
fn: function(args) {
if (args.length < 2) throw new Error('Not enough args to contains()');
return checkStr(args[0]).toLowerCase().includes(checkStr(args[1]).toLowerCase());
}
},
atkspdmod: {
type: 'number',
fn: function(args) {
if (args.length < 1) throw new Error('Not enough args to atkSpdMod()');
switch (checkNum(args[0])) {
case 2:
return 3.1;
case 1:
return 2.5;
case 0:
return 2.05;
case -1:
return 1.5;
case -2:
return 0.83;
}
if (args[0] <= -3) return 0.51;
if (args[0] >= 3) return 4.3;
throw new Error('Invalid argument to atkSpdMod()');
}
if (args[0] <= -3) return 0.51;
if (args[0] >= 3) return 4.3;
throw new Error('Invalid argument to atkSpdMod()');
}
};
// the compiler itself
const compileQueryExpr = (function() {
// tokenize an expression string
function tokenize(exprStr) {
exprStr = exprStr.trim();
const tokens = [];
let col = 0;
function pushSymbol(sym) {
tokens.push({ type: 'sym', sym });
col += sym.length;
}
while (col < exprStr.length) {
// parse fixed symbols, like operators and stuff
switch (exprStr[col]) {
case '(':
case ')':
case ',':
case '&':
case '|':
case '+':
case '-':
case '*':
case '/':
case '^':
case '=':
pushSymbol(exprStr[col]);
continue;
case '>':
pushSymbol(exprStr[col + 1] === '=' ? '>=' : '>');
continue;
case '<':
pushSymbol(exprStr[col + 1] === '=' ? '<=' : '<');
continue;
case '!':
pushSymbol(exprStr[col + 1] === '=' ? '!=' : '!');
continue;
case ' ': // ignore extra whitespace
++col;
continue;
}
if (exprStr.slice(col, col+2) === "?=") {
pushSymbol("?=");
continue;
}
// parse a numeric literal
let m;
if ((m = /^\d+(?:\.\d*)?/.exec(exprStr.substring(col))) !== null) {
tokens.push({ type: 'num', value: parseFloat(m[0]) });
col += m[0].length;
continue;
}
// parse a string literal
if ((m = /^"([^"]+)"/.exec(exprStr.substring(col))) !== null) { // with double-quotes
tokens.push({ type: 'str', value: m[1] });
col += m[0].length;
continue;
}
if ((m = /^'([^']+)'/.exec(exprStr.substring(col))) !== null) { // with single-quotes
tokens.push({ type: 'str', value: m[1] });
col += m[0].length;
continue;
}
// parse an identifier or boolean literal
if ((m = /^\w[\w\d%]*/.exec(exprStr.substring(col))) !== null) {
switch (m[0]) {
case 'true':
tokens.push({ type: 'bool', value: true });
col += 4;
continue;
case 'false':
tokens.push({ type: 'bool', value: false });
col += 5;
continue;
}
tokens.push({ type: 'id', id: m[0] });
col += m[0].length;
continue;
}
// if we reach here without successfully parsing a token, it's an error
throw new Error(`Could not parse character "${exprStr[col]}" at position ${col}`);
}
tokens.push({ type: 'eof' });
return new TokenList(tokens);
// static type check
function staticCheck(expType, term) {
if (expType === 'any' || expType === term.type) {
return true;
}
throw new Error(`Expected ${expType}, but got ${term.type}`);
}
// expression terms
class Term {
constructor(type) {
this.type = type;
}
// parse tokens into an ast
function takeDisj(tokens) {
const left = takeConj(tokens);
if (tokens.here.type === 'sym' && tokens.here.sym === '|') {
tokens.advance();
const right = takeDisj(tokens);
return (i, ie) => checkBool(left(i, ie)) || checkBool(right(i, ie));
}
return left;
resolve(item, itemExt) {
throw new Error('Abstract method!');
}
}
class LiteralTerm extends Term {
constructor(type, value) {
super(type);
this.value = value;
}
function takeConj(tokens) {
const left = takeCmpEq(tokens);
if (tokens.here.type === 'sym' && tokens.here.sym === '&') {
tokens.advance();
const right = takeConj(tokens);
return (i, ie) => checkBool(left(i, ie)) && checkBool(right(i, ie));
}
return left;
resolve(item, itemExt) {
return this.value;
}
}
class BoolLitTerm extends LiteralTerm {
constructor(value) {
super('boolean', value);
}
}
class NumLitTerm extends LiteralTerm {
constructor(value) {
super('number', value);
}
}
class StrLitTerm extends LiteralTerm {
constructor(value) {
super('string', value);
}
}
class BinaryOpTerm extends Term {
constructor(type, leftType, left, rightType, right) {
super(type);
staticCheck(leftType, left);
staticCheck(rightType, right);
this.left = left;
this.right = right;
}
function takeCmpEq(tokens) {
const left = takeCmpRel(tokens);
if (tokens.here.type === 'sym') {
switch (tokens.here.sym) {
case '=': {
tokens.advance();
const right = takeCmpEq(tokens);
return (i, ie) => {
const a = left(i, ie), b = right(i, ie);
if (typeof a !== typeof b) return false;
switch (typeof a) {
case 'number':
return Math.abs(left(i, ie) - right(i, ie)) < 1e-4;
case 'boolean':
return a === b;
case 'string':
return a.toLowerCase() === b.toLowerCase();
}
throw new Error('???'); // wut
};
}
case '!=': {
tokens.advance();
const right = takeCmpEq(tokens);
return (i, ie) => {
const a = left(i, ie), b = right(i, ie);
if (typeof a !== typeof b) return false;
switch (typeof a) {
case 'number':
return Math.abs(left(i, ie) - right(i, ie)) >= 1e-4;
case 'boolean':
return a !== b;
case 'string':
return a.toLowerCase() !== b.toLowerCase();
}
throw new Error('???'); // wtf
};
}
case '?=': {
tokens.advance();
const right = takePrim(tokens);
return (i, ie) => {
const a = left(i, ie), b = right(i, ie);
if (typeof a !== typeof b) return false;
switch (typeof a) {
case 'number':
return Math.abs(left(i, ie) - right(i, ie)) < 1e-4;
case 'boolean':
return a === b;
case 'string':
return a.toLowerCase().includes(b.toLowerCase());
}
throw new Error('???'); // wtf
};
}
}
}
return left;
resolve(item, itemExt) {
return this.apply(this.left.resolve(item, itemExt), this.right.resolve(item, itemExt));
}
function takeCmpRel(tokens) {
const left = takeSum(tokens);
if (tokens.here.type === 'sym') {
switch (tokens.here.sym) {
case '<=': {
tokens.advance();
const right = takeCmpRel(tokens);
return (i, ie) => checkNum(left(i, ie)) <= checkNum(right(i, ie));
}
case '<': {
tokens.advance();
const right = takeCmpRel(tokens);
return (i, ie) => checkNum(left(i, ie)) < checkNum(right(i, ie));
}
case '>': {
tokens.advance();
const right = takeCmpRel(tokens);
return (i, ie) => checkNum(left(i, ie)) > checkNum(right(i, ie));
}
case '>=': {
tokens.advance();
const right = takeCmpRel(tokens);
return (i, ie) => checkNum(left(i, ie)) >= checkNum(right(i, ie));
}
}
}
return left;
apply(a, b) {
throw new Error('Abstract method!');
}
}
class LogicalTerm extends BinaryOpTerm {
constructor(left, right) {
super('boolean', 'boolean', left, 'boolean', right);
}
}
class ConjTerm extends LogicalTerm {
apply(a, b) {
return a && b;
}
}
class DisjTerm extends LogicalTerm {
apply(a, b) {
return a || b;
}
}
class EqualityTerm extends BinaryOpTerm {
constructor(left, right) {
super('boolean', 'any', left, 'any', right);
}
}
class EqTerm extends EqualityTerm {
apply(a, b) {
return a === b;
}
}
class NeqTerm extends EqualityTerm {
apply(a, b) {
return a !== b;
}
}
class ContainsTerm extends BinaryOpTerm {
constructor(left, right) {
super('boolean', 'string', left, 'string', right);
}
function takeSum(tokens) {
const left = takeProd(tokens);
if (tokens.here.type === 'sym') {
switch (tokens.here.sym) {
case '+': {
tokens.advance();
const right = takeSum(tokens);
return (i, ie) => checkNum(left(i, ie)) + checkNum(right(i, ie));
}
case '-': {
tokens.advance();
const right = takeSum(tokens);
return (i, ie) => checkNum(left(i, ie)) - checkNum(right(i, ie));
}
}
}
return left;
apply(a, b) {
return a.toLowerCase().includes(b.toLowerCase());
}
}
class InequalityTerm extends BinaryOpTerm {
constructor(left, right) {
super('boolean', 'any', left, 'any', right);
}
function takeProd(tokens) {
const left = takeExp(tokens);
if (tokens.here.type === 'sym') {
switch (tokens.here.sym) {
case '*': {
tokens.advance();
const right = takeProd(tokens);
return (i, ie) => checkNum(left(i, ie)) * checkNum(right(i, ie));
}
case '/': {
tokens.advance();
const right = takeProd(tokens);
return (i, ie) => checkNum(left(i, ie)) / checkNum(right(i, ie));
}
}
}
return left;
apply(a, b) {
checkComparable(a);
checkComparable(b);
return this.compare(a, b);
}
function takeExp(tokens) {
const left = takeUnary(tokens);
if (tokens.here.type === 'sym' && tokens.here.sym === '^') {
tokens.advance();
const right = takeExp(tokens);
return (i, ie) => checkNum(left(i, ie)) ** checkNum(right(i, ie));
}
return left;
compare(a, b) {
throw new Error('Abstract method!');
}
}
class LeqTerm extends InequalityTerm {
compare(a, b) {
return a <= b;
}
}
class LtTerm extends InequalityTerm {
compare(a, b) {
return a < b;
}
}
class GtTerm extends InequalityTerm {
compare(a, b) {
return a > b;
}
}
class GeqTerm extends InequalityTerm {
compare(a, b) {
return a >= b;
}
}
class ArithmeticTerm extends BinaryOpTerm {
constructor(left, right) {
super('number', 'number', left, 'number', right);
}
}
class AddTerm extends ArithmeticTerm {
apply(a, b) {
return a + b;
}
}
class SubTerm extends ArithmeticTerm {
apply(a, b) {
return a - b;
}
}
class MulTerm extends ArithmeticTerm {
apply(a, b) {
return a * b;
}
}
class DivTerm extends ArithmeticTerm {
apply(a, b) {
return a / b;
}
}
class ExpTerm extends ArithmeticTerm {
apply(a, b) {
return a ** b;
}
}
class UnaryOpTerm extends Term {
constructor(type, inType, inVal) {
super(type);
staticCheck(inType, inVal);
this.inVal = inVal;
}
function takeUnary(tokens) {
if (tokens.here.type === 'sym') {
switch (tokens.here.sym) {
case '-': {
tokens.advance();
const operand = takeUnary(tokens);
return (i, ie) => -checkNum(operand(i, ie));
}
case '!': {
tokens.advance();
const operand = takeUnary(tokens);
return (i, ie) => !checkBool(operand(i, ie));
}
}
}
return takePrim(tokens);
resolve(item, itemExt) {
return this.apply(this.inVal.resolve(item, itemExt));
}
function takePrim(tokens) {
switch (tokens.here.type) {
case 'num': {
const lit = tokens.here.value;
tokens.advance();
return (i, ie) => lit;
}
case 'bool': {
const lit = tokens.here.value;
tokens.advance();
return (i, ie) => lit;
}
case 'str': {
const lit = tokens.here.value;
tokens.advance();
return (i, ie) => lit;
}
case 'id':
const id = tokens.here.id;
tokens.advance();
if (tokens.here.type === 'sym' && tokens.here.sym === '(') { // it's a function call
tokens.advance();
const argExprs = [];
if (tokens.here.type !== 'sym' || tokens.here.sym !== ')') {
arg_iter: // collect arg expressions, if there are any
while (true) {
argExprs.push(takeDisj(tokens));
if (tokens.here.type === 'sym') {
switch (tokens.here.sym) {
case ')':
tokens.advance();
break arg_iter;
case ',':
tokens.advance();
continue;
}
}
throw new Error(`Expected "," or ")", but got ${JSON.stringify(tokens.here)}`);
}
}
const func = itemQueryFuncs[id.toLowerCase()];
if (!func) throw new Error(`Unknown function: ${id}`);
return (i, ie) => {
const args = [];
for (let k = 0; k < argExprs.length; k++) args.push(argExprs[k](i, ie));
return func(args);
};
} else { // not a function call
const prop = itemQueryProps[id.toLowerCase()];
if (!prop) throw new Error(`Unknown property: ${id}`);
return prop;
}
case 'sym':
if (tokens.here.sym === '(') {
tokens.advance();
const expr = takeDisj(tokens);
if (tokens.here.type !== 'sym' || tokens.here.sym !== ')') throw new Error('Bracket mismatch');
tokens.advance();
return expr;
}
break;
}
throw new Error(tokens.here.type === 'eof' ? 'Reached end of expression' : `Unexpected token: ${JSON.stringify(tokens.here)}`);
apply(x) {
throw new Error('Abstract method!');
}
}
class NegTerm extends UnaryOpTerm {
constructor(inVal) {
super('number', 'number', inVal);
}
// full compilation function, with extra safety for empty input strings
return function(exprStr) {
const tokens = tokenize(exprStr);
return tokens.tokens.length <= 1 ? null : takeDisj(tokens);
};
})();
apply(x) {
return -x;
}
}
class InvTerm extends UnaryOpTerm {
constructor(inVal) {
super('boolean', 'boolean', inVal);
}
apply(x) {
return !x;
}
}
class FnCallTerm extends Term {
constructor(fn, argExprs) {
super(fn.type);
this.fn = fn;
this.argExprs = argExprs;
}
resolve(item, itemExt) {
const argVals = [];
for (const argExpr of this.argExprs) {
argVals.push(argExpr.resolve(item, itemExt));
}
return this.fn.fn(argVals);
}
}
class PropTerm extends Term {
constructor(prop) {
super(prop.type);
this.prop = prop;
}
resolve(item, itemExt) {
return this.prop.resolve(item, itemExt);
}
}